Researchers trial AI coach for diabetes
· Medical Xpressedited by Lisa Lock, reviewed by Andrew Zinin
Lisa Lock
Scientific Editor
Meet our editorial team
Behind our editorial process
Andrew Zinin
Chief Editor
Meet our editorial team
Behind our editorial process Editors' notes
This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:
fact-checked
trusted source
proofread
The GIST Add to Preferred Sources
For many people with type 2 diabetes, the challenge isn't understanding what they need to do to stay well. It's doing it consistently. Managing the condition often requires a long-term commitment to healthier eating patterns, regular exercise and medication. But unlike some illnesses, diabetes can progress quietly. Missing a dose or skipping exercise may not cause immediate consequences, making it easy for healthy intentions to slip amid the demands of work, family and everyday life.
Researchers at the Auckland Bioengineering Institute (ABI) at Waipapa Taumata Rau, University of Auckland, are developing an AI-powered digital health coach called DiNa, designed to support people between medical appointments and help them stay on track with goals set alongside their health care team. Dr. Gonzalo Maso Talou leads the team for the project, which is part of ABI's broader digital twin program.
DiNa is currently being tested in a clinical trial involving people with type 2 diabetes. "DiNa is not about curing diabetes," says Maso Talou. "What we're trying to do is increase people's chances of adhering to the treatment plan they've developed with their clinician."
The need is significant. According to Health New Zealand's Virtual Diabetes Register, more than 365,000 people were living with diabetes in Aotearoa in 2025, and numbers have increased by more than 40% since 2011. Diabetes disproportionately affects Māori, Pacific and Indian communities.
The current study will follow 10 participants with type 2 diabetes over three months. Half are using DiNa, while half are using an existing diabetes-management tool, allowing researchers to compare outcomes.
The team, which includes Ph.D. student Kekayan Nanthakumar, is interested in a simple but important question: Can personalized, ongoing support help people maintain healthy habits more effectively?
To answer that, DiNa draws information from multiple sources, including meal records, medication tracking, smartphone activity data and continuous glucose monitors (CGMs). Rather than asking users to carefully record every detail themselves, the system aims to make monitoring as easy as possible.
Users can simply photograph their meals, with artificial intelligence identifying foods and estimating calories, carbohydrates, protein and other nutritional information. If information is missing, users can correct or add it using voice commands.
"It removes a lot of the burden of manually keeping food diaries," says Maso Talou. "That's important because traditional tracking is time-consuming and people often forget details or stop recording altogether."
For exercise, DiNa can draw on information already collected by smartphones and wearable devices. Users can also quickly record whether they have taken their prescribed medications.
But DiNa's most distinctive feature is conversation.
Instead of functioning as a traditional health app, the system allows users to chat or speak with an AI coach through a voice-only interface. Early DiNa prototypes involved talking heads on screen, but participants didn't necessarily warm to that, Maso Talou says.
"We originally thought a realistic virtual person would make the experience better. What we found was many users preferred just talking by voice. It was less demanding and felt more natural." The result is an interaction that feels more like a phone conversation than a video call.
DiNa might help someone work out the best time of day to exercise, identify barriers that are preventing them from achieving goals or suggest ways to adapt plans when life gets in the way. The system can provide positive encouragement when things are going well and help users problem-solve when they are not.
Importantly, the digital coach is not designed to replace clinicians. The information it provides comes from curated New Zealand diabetes resources developed with clinical partners, and it directs users back to health care professionals whenever clinical advice is needed, Maso Talou says.
"There can be months between appointments. People often have questions or uncertainties during that time. DiNa can help people better understand their condition and the advice they've already received."
Participants in the study are already responding positively to the technology. "One of the first things people told us was how much easier it made monitoring. Many were already familiar with tracking food and medication. They liked that the system reduced that workload and allowed them to focus on getting on with their lives."
Beyond behavior change, researchers are working toward a more ambitious goal: personalized digital twins. Digital twins are computer models that simulate how an individual's body responds to factors such as food, exercise and medication. Early versions are already able to model glucose responses using information about carbohydrates, fats and proteins.
Eventually, that could allow users to ask highly personalized questions before making decisions. For example, someone could compare two meal options and see which is more likely to cause a glucose spike; they could ask their coach to generate healthy recipes based on their glucose levels and the ingredients they already have at home; or they could explore how changing a meal might affect their blood sugar levels.
"Two people can respond very differently to exactly the same meal," says Maso Talou, who is co-lead on the digital twin project with Auckland Bioengineering Institute colleague Dr. Finbar Argus. "Factors such as sex, metabolism, activity levels and other physiological differences all play a role. The digital twin helps us move toward truly personalized advice."
Researchers believe the technology could become a powerful educational tool, helping people better understand how their own bodies respond to different behaviors and empowering them to make informed choices.
The initial three-month study is the first of three planned trial rounds, with larger studies scheduled next year. Data collected through those trials will also help researchers refine the underlying digital twin technology. If the results are positive, the team hopes DiNa could eventually be integrated into health care systems and made widely available.
For Maso Talou, the long-term goal is straightforward. "We want to help people make sustainable lifestyle changes and stay engaged with their care," he says. "If we can make it easier for people to follow the plans they've already developed with their health care team, we have an opportunity to improve health outcomes and quality of life."
Key medical concepts
Clinical categories
EndocrinologyHealthy livingNutrition & Healthy eatingFitness & Physical activityCommon illnesses & Prevention Provided by University of Auckland Who's behind this story?
Lisa Lock
BA art history, MA material culture. Former museum editor, paramedic, and transplant coordinator. Editing for Science X since 2021. Full profile →
Andrew Zinin
Master's in physics with research experience. Long-time science news enthusiast. Plays key role in Science X's editorial success. Full profile →
Citation: Researchers trial AI coach for diabetes (2026, November 13) retrieved 13 November 2026 from https://medicalxpress.com/news/2026-11-trial-ai-diabetes.html This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.